外文翻译--基于案例的液压动力回路设计 英文版【优秀】.pdf -- 10 元

Y.WuEd.ICHCC2011,CCIS163,pp.269–275,2011.©SpringerVerlagBerlinHeidelberg2011CaseBasedDesignforHydraulicPowerCircuitChiManVong,PakKinWong,WengFaiIp,andZhiXinYangFacultyofScienceandTechnology,UniversityofMacau,Macau{cmvong,fstpkw,andyip,zxyang}umac.moAbstract.ThispaperdescribesthedesignandimplementationofanautomatichydrauliccircuitdesignsystemusingcasebasedreasoningCBRasoneofthesuccessfulartificialintelligenceparadigms.Thedomainofcasebasedreasoningandhydrauliccircuitdesignarebrieflyreviewed.ThenaproposedmethodologyinautomaticcircuitdesignandlearningwiththeuseofCBRisdescribed.FinallyanapplicationexamplehasbeenselectedtoillustratetheusefulnessofapplyingCBRinindustrialhydrauliccircuitdesignwithlearning.KeywordsCasebasedreasoningCBR,adaptationcase,hydrauliccircuitdesign.1IntroductionTheuseofcomputersinengineeringdesignhasbecomeamajortrendinindustry.Today,differentcommercialautomaticcomputeraideddesignCADsoftwareisavailabletoautomatethedesignprocessinmanyengineeringapplications.However,CADsoftwareinhydraulicsystemdesignisnotasprominentasinmanyotherareasofengineeringdesign.Thisismainlyduetothecomplexityofhydraulicanalysisandlackofagreementofthemostappropriateapproachtothedesignprocess.Inrecentyears,manyresearchesonintelligentCADorexpertsystemsforhydrauliccircuitshavebeenfoundintheliterature.MostoftheCADsystemsarebuiltfromproductionrules1fordesignknowledgerepresentationorintegratedrulebasedandobjectorientedtechnology2forreducingthecomplexityinhydraulicsubcircuitandcomponentrepresentation.Althoughtheapproachesareeffective,theacquisitionandmaintenanceofrulesaretheproblemsfacingbynotonlythesoftwareengineersbutalsothedesignersusingthesoftware.Moreover,staticlearning1isanotherissueoftraditionalrulebasedsystems.Toresolvetheproblemsinheritedfromconventionalrulebasedexpertsystems,theAIcommunityproposedanotherreasoningparadigmcalledcasebasedreasoningCBR.CBRsupportslearninginthewaythatnewknowledgecanbeappendedtotheknowledgebasewithoutwiderrecompilationofthesystem.ThisisoneofmainadvantagesofCBRthatmaintenanceofknowledgetakesmuchlesstime.ThispaperstudiestheapplicationofCBRinhydraulicsystemdesignandaprototypeautomatichydrauliccircuitdesignsystemhasbeendevelopedtoverifythisproposedmethodology.1Whenevertherulesareupdated,thewholesystemhastoberecompiled.270C.Vongetal.2Review2.1CaseBasedReasoningCBR3isamethodologythatallowsdiscoveringanalogiesbetweenacurrentworkingsituationandpastexperiencesreferencecases.CBRmakesdirectuseofpastexperiencestosolveanewproblembyrecognizingitssimilaritywithaspecificknownproblemandbyapplyingitssolutiontofindasolutionforthecurrentsituation.CBRhasbeenusedtodevelopmanysystemsappliedinavarietyofdomains4,5,6,7,8,9,10,includingmanufacturing,design,law,medicine,andplanning.BasicallyCBRisconstitutedwithfourREs3RETRIEVE––retrievesimilarpastcasematchedagainstcurrentproblemREUSE––reusetosolvecurrentproblembasedonsolutionofpastcaseREVISE––revisethepastsolutionifanycontradictionoccurswhenappliedtocurrentproblemRETAIN––retainthefinalsolutionalongwiththeproblemasacaseifthecaseisusefulinthefutureWhentheuserinputsaproblem,theproblemisinterpretedandconvertedasanewcaseintothespecificformatofthereasoningsystem.ThentheconvertednewcaseentersthestageofRETRIEVALwherethenewcaseismatchedagainstthepreviouscasesinthecaselibraryofthereasoningsystem.IntheretrievalstageofCBRusuallyasimplesimilarityfunctionisemployedtofindthenearestneighborforthecurrentproblemfromthereferencecases.Theformulaislistedin1wherewiistheimportanceofdimensioni,fiIandfiRarethevaluesforfeaturefiintheinputandretrievedcases,respectively.Forsymbolicvaluesoff,0if1ifIRIRiiiiIRiiffffff−≠.211,,nIRiiiIRiniiwffEffWw−.12.2HydraulicCircuitDesignHydraulicpowerisoneofpowertransmissionsystemsandcontrol.Itconvertsmechanicalenergytohydraulicenergyforproducingusefulworksuchaspressingorlifting.Themaintaskofhydraulicpowersystemdesigniscircuitdesign.ThegeneralprocedureisshownasfollowsI.Thecircuitisdesignedaccordingtotheinformationprovidedbytheclientsuchasmaximumoperatingpressure,maximumload,speedofactuator,dutycycleandapplication,etc.II.Thesizesofthelinearorrotaryactuatorsaredeterminedaccordingtothemaximumoperatingpressure,maximumloadandloaddisplacement.III.Convertthecalculatedactuatorsizesintostandardsizes.CaseBasedDesignforHydraulicPowerCircuit271IV.Thesystemparameterssuchashydraulicoilflowrate,pressurechangeover,etc,aredetermined.V.Thesuitableactuatorsubcircuits,pumpandpumpsubcircuitisselectedbasedonthesystemparametersandthedesignspecifications.VI.Otherhydrauliccomponentsusedinthecircuitarefinallyselected.3ApplyingCBRinIndustrialHydraulicCircuitDesignInsection2.2stepV,hydraulicengineersareusuallyaccustomedtomodifyanexistingcircuitdesignintoanewonefordifferentsituationanduse.Itisbecausehydrauliccircuitdesignisusuallysimilarevenfordifferentsituation,sohydraulicengineershavetomanuallylookthroughmanyexistingeffectivecircuitsandthenselectasimilarandsuitableoneandperformmodification.Theprocessisrepetitiveandtediousinthestageofretrieving,becauseengineershavetoreviewthecircuitsonebyone.However,theprocessoftheretrievinganexistingcircuitandadaptingittofitbetterthecurrentsituationcanbestronglysupportedbyCBR.Iftheexistingcircuitsarecollectedinacomputerdatabasecaselibrary,andeachcircuitisstoredalongwithitsfunctionalandapplicationspecificrequirementsoftheoutputs,theseparameterscanserveasthecasecircuitdesignindexes.Wheneverthehydraulicengineerwantstoretrieveapastcasefromthecaselibrary,hejustneedstoinputthefunctionalandapplicationspecificrequirements,andCBRuses1tocalculatethemostsimilarpastcase.Iftheengineerisnotsatisfiedwiththerecommendedcase,thenthenextmostsimilarpastcasewouldbeshown.Thiscouldbedonebecausethecasesarealreadyrankedwithdifferentsimilaritylevelinthecalculationofcasesimilarity.Afterthat,theengineercouldadaptormodifytheretrievedcasemanuallybytheaboveprocedureorbyapplyingtheadaptationrulessuppliedbyCBR.Thoseadaptationrulesarespecificproductionrulescapturedfromexpertsorfromtheengineersownexperience.Finallytheengineercandecideifthecaseisgoodenoughtostoreintothecaselibraryforfutureuse.Hydrauliccircuitdesignsdifferbynotonlythecircuitdiagrambutalsothefunctionalattributesalongwiththem.Forcircuitadaptation,ifsomecomponentsarereplaced,thentheattributeswillalsobemodifiedbyinsertingordeletingsomeofthem.CBRenablesstructuralmodificationofcases,sothatattributescanbeaddedordeletedaccordingly.Forexample,anengineerretrievesanpastcaseandperformsmodificationonthecircuitdesign,thenthenumberofattributeswouldbechangedaccordingto,whichsubcircuitsarerevisedbyaddingpredefinedattributesordeletingunnecessaryattributesinthesubcircuits.4ApplicationExampleThesystemimplementedisabletorecommendmostsimilarcircuitdesignbasedonthecircuitspecification.TheworkingenvironmentandfrontenduserinterfaceofthecircuitdesignsystemareshowninFig.1.Thelearningabilityofthesystemisillustratedinthispartwiththeaidofanexample.Table1showspartialattributesofthecaserepresentationforahydraulicsubcircuitshowninFig.2.272C.Vongetal.Fig.1.WorkingenvironmentandfrontenduserinterfaceofthecircuitdesignsystemTable1.PartialattributesofahydraulicsubcircuitDrawingnameVar_1Max.Flow33L/minMax.Pressure630BarVariationstepsofspeed2Variationstepsofpressure1Wheneverapastsimilarcaseisretrieved,thecaseisadaptedinordertoreuseitforcurrentsituation.However,noteverycaseisadaptablebythesystem,astheadaptationrulesetofanysystemisalwaysincomplete.Atthismoment,theuserinterventionisnecessarytocompensatetheinabilityofadaptationofthesystem.Theuserswilladaptthecaseusinghisdomainexpertise.Inordertolearnthedomainknowledgefromtheuser,theoperationperformedbytheuserisrecordedbymeansofansweringquestionsinstep.Whenauserwantstoadaptacase,thesystemwillasktheuserwhichkindofoperationstoperform.Oncetheuserhaschosentheoperation,thecorrespondingactionsaresuppliedtotheusertochoose.Inrecordingtheoperations,thesystemcanlearnfromtheusertheadaptationknowledge.Thelearntknowledgeisencapsulatedasacasecalledadaptationcasebecauseitisusedtoguideadaptation.Subsequently,whensimilarproblemcasearises,thesystemwillbecomecapabletohandletheadaptationbyreferringtotheadaptationcase.